|OBJECTIVE & APPROACH|
Denitrification Modeling Across Terrestrial, Freshwater and Marine Systems
November 28-30, 2006
At The Institute of Ecosystem Studies, Millbrook, New York, USA
POSTER TITLES & ABSTRACTS (17)
Controls on N accumulation
and loss in arctic tundra ecosystems
History and fate of nitrate
from isotopes and other environmental tracers
J. K. Böhlke
Stream nitrate uptake and
denitrification rates determined from15N addition experiments:
Preliminary results from the LINX II study
Pat Mulholland, Bob Hall, Steve Hamilton, Bruce Peterson, Jen Tank, Linda Ashkenas, Lee Cooper, Cliff Dahm, Walter Dodds, Stuart Findlay, Stan Gregory, Nancy Grimm, Sherri Johnson, Bill McDowell, Judy Meyer, Geoff Poole, Maury Valett, Jack Webster[abstract] [pdf]
In order to estimate gaseous N volatilization and emissions of environmental relevant N trace gases (NO and N2O) from terrestrial ecosystems in Europe we established a detailed spatial database on land use, soil properties, agricultural management and crops, N deposition and climate. Data were obtained from various sources and were harmonized for defined grid cell sizes of 10 km2 (agriculture) or 50 km2 (forest). The DNDC (agriculture) or Forest-DNDC (forests) models were linked to the GIS database for calculating European wide emission inventories. Depending on the individual simulation year N2O emissions from arable soils were approx. 2-3 times higher as from forest soils (forests EU 15: 55-65 kt N), whereas soil NO emissions were in the same range for agricultural and forest soils. Total simulated gaseous N losses (incl. N2) from agricultural and forest ecosystems in EU 15 where in a range of 400 - 700 kt N. Compared to simulated N leaching (simulation depth 0.3 m) of approx. 4000 kt N this seems to be a rather low number, indicating that the production of N2 via denitrification may not adequately be described or calibrated in DNDC. On the other hand this alaso shows that a simulation depth of 0.3 m is not sufficient for estimating total gaseous N losses, since one needs to assume that most of the leached N will be further denitrified in deeper soil layers.
Furthermore, we also investigated possible feedbacks of predicted future climate change on forest soil NO and N2O emissions in Europe. Two climate scenarios were used: one representing a 10-year period of present day climate (1991-2000), the other a nine-year period for future climate conditions (2031-2039). The scenarios were used to drive the GIS-coupled Forest-DNDC, which has currently been tested for its predicting capability for soil N trace gas emissions for various sites across Europe. The model results show a complex, spatially differentiated pattern of changes in future N2O and NO emissions from forest soils across Europe, which were caused by the combined effect of changes in precipitation and temperature. Overall, the model predicted that N2O emissions will on average decrease by 6%. This decrease was mainly due to a shift in the N2O:N2 ratio driven by enhanced denitrification. NO emissions were found to increase by 9%. The increases in NO emissions were mainly due to increases in temperature. Only for the regions where soil moisture was predicted to markedly increase or suffer from water stress during the vegetation period, a reduction of NO emissions was simulated. However, further process studies are needed to confirm model predictions and to improve the parameterization of N2 production.
controls of aquatic denitrification: Examples from estuaries, reservoirs
Cornwell, J.C., M.S. Owens and T.M. Kana
Using the MIMS approach to measuring denitrification (via gas ratios) we have examined the environmental controls of denitrification in a number of different aquatic environments. Although overall sediment metabolic rate and the rate of nitrate supply from nitrification or overlying water are ultimately the proximal controls of denitrification at the microbial scale, a number of other factors are key to the the process of denitrification. In freshwater and marine systems, the same general controls apply.
In the Chesapeake Bay, denitrification is generally controlled by a) watershed nitrate supply in the upper tidal freshwater/oligohaline environments, b) rates of nitrification, particularly regarding the influence of oxygen at the sediment-water interface, and c) competition for ammonium between nitrifiers and benthic microalgae. In illuminated estuarine sediments from temperate (Long Island, Maryland Coastal Bays) and subtropical (Florida Bay) environments, denitrification occurs at higher rates in the dark than in the light. Denitrification at the surface of river sediments (Indiana) is controlled by overall rates of sediment respiration in the dark, and by competition for substrate with benthic microalgae in the light. Finally, we show data regarding the time of recovery of coupled nitrification-denitrification after aeration of anoxic-sulfidic sediments from the Chesapeake Bay.
Modeling aquatic denitrification, with its many different environmental controls, remains a significant challenge. Beyond the controls described above, physical/biological effects such as bioirrigation and bioturbation can also be important. An important question that remains to be answered is whether measurements of denitrification in diverse environments is sufficient for model calibration, or whether targeted experiments are more appropriate.
The DAYCENT biogeochemical model has recently been adopted by the EPA to estimate N2O emissions from agricultural soils for the US Greenhouse Gas Inventory. At the sub-county level, native vegetation and modern land cover were simulated and outputs for 1990-2004 were compiled for N gas and NO3 leaching losses. Three sets of simulations were performed for each county, one for the land area where small grains and row crops are the dominant agricultural practice, one to represent where pastures and hay dominate, and one for rangeland. Meteorological data required to drive DAYCENT were acquired from DAYMET, an algorithm that uses weather station data and accounts for topography to predict daily temperature and precipitation at 1 km2 resolution. Soils data were acquired from STATSGO. All the DAYMET cells that are covered with land where row/small grain cropping is the dominant agricultural land use were identified and we selected the DAYMET cell with the median annual precipitation to represent all of this type of land area. Then, we selected the soil texture class that intersected the most land area used for row/small cropping. Similar procedures were performed for areas where pasture/hay cropping and rangeland are the dominant agricultural practices. N loss vectors were calculated by accounting for the present day land distribution of cropped soils and non-cropped soils, with the latter assumed to be covered by potential native vegetation. Area weighted mean county level N2 and NO3 loss maps show that denitrification is highest in intensively cropped fine textured soils (e.g., along the Mississippi, some counties in Texas, some irrigated western counties) and in some northern counties where subsoils remain frozen in spring. Leaching and total N losses where greatest from coarse textured soils (e.g., the southeast) and from intensively cropped loam soils in the Midwest. Sub-county N loss maps for the Merrimack and Illinois watersheds will also be presented.
In sharp contrast to the sediment denitrification measured in Narragansett Bay during the 1970s, 1990s, and in 2005, there was a net uptake of N2 by the sediments in July and August of 2006. N2 uptake was measured in each of the triplicate cores at four stations in the mid and upper portion of the bay using membrane inlet mass spectrometry (MIMS). Net N2-fixation rates ranged from 60 to 250 µmol N2-N m-2 h-1 which were higher than the net denitrification rates measured during 2005. We believe this dramatic shift is due to various ecological changes which have occurred over the last three decades in Narragansett Bay. An increase in the annual mean water temperature (+1.7ºC) of the bay can be linked to a dramatic drop in the standing crop of water column chlorophyll as the system has shifted from one characterized by a dominant winter-spring bloom to one supported by more ephemeral and less intense summer-fall blooms. Intense grazing by ctenophores in the spring and summer of 2006 (but not 2005) virtually eliminated the summer copepods. This combination may have resulted in more organic consumption in the water column, leaving the summer sediments with very little available carbon and nitrogen.
variability of nitrogen dynamics in a freshwater reservoir
The Eagle creek reservoir (ECR) is a stream-fed freshwater reservoir which, over the years, has evolved into a nutrient-rich system manifested in recent episodes of algal blooms. Although the productivity of aquatic ecosystems has generally been linked to P availability, several studies have also demonstrated that, in eutrophic systems, nitrogen (N) availability is often the factor controlling primary productivity.
ECR comprises two distinct sections: a shallow (average depth, 2.5 m) and well mixed section, and a deep (up to 15 m) and generally stratified section. Nitrogen mass balance, based on mineral N (NO3- and NH4+) concentration in ECR tributaries and outlets in 2004, showed that 42 % of the annual N loading was removed. Assuming that denitrification was the predominant N sink, the corresponding denitrification rate averaged 72.5 ± 46 mg N m-2 d-1. In addition to the mass balance approach, mineral N was also monitored at different depths and locations throughout the reservoir. In the shallow section of the reservoir, NO3- concentration within the reservoir mirrored that in the inlets and exhibited similar temporal patterns. In contrast, although NO3- input into the deep section was 3-4 fold higher (mean: 5.9 mg N L-1), reservoir NO3- concentration varied independently of, and was always < 30 % inlet concentration. The main NO3- sink in ECR appears to reside, not in the shallow, but in the deep section of the reservoir. These results contrasted with those obtained from denitrification enzyme assays (summer and late fall) which showed limited spatial and temporal variation in the denitrification capacity (120 mg N m-2 d-1) of ECR sediment. Concentration profiles in the deep section also indicated a flux of NH4+ from the sediment into the water column. An inverse relationship between dissolved O2 and NH4+ flux was observed, suggesting that dissimilatory nitrate reduction to ammonia (DNRA) may be active in ECR sediment. The significance of DNRA is being evaluated through field studies combining perfusion of 15NO3- and diffusive equilibrium probes. These results will be presented and discussed along with events affecting water column mixing regime.
ammonium oxidation in tidal marshes
Denitrification is recognized as an important microbial process, which removes inorganic nitrogen from marsh sediments, by reducing nitrate to dinitrogen gas, which is consequently exported to the atmosphere. During the last 5 years another dinitrogen-producing microbial process, Anaerobic Ammonium Oxidation (anammox) has been discovered naturally occurring in various marine environments. The importance of anammox in marine environments varies and the natural physico-chemical and microbial controls are still largely unknown. This study investigated anammox and denitrification activity in salt marsh sediments from the marshes associated with the Plum Island Ecosystem LTER site, MA. Potential N2-production was measured in various vegetation zones over a salinity gradient ranging from Spartina spp. dominated salt marsh to the oligohaline Typha spp. dominated marsh. Anaerobic ammonium oxidation was present in most salt marsh habitats, but at rates orders of magnitude lower than the rates of denitrification. Anammox accounts for less than 3% of the total N2-production and is insignificant as a nitrogen exporting process in salt marsh sediment. The impact of increased nitrate availability on denitrification and anammox activity was studied in a fertilized tidal creek and a marsh with naturally high nitrate abundance. Denitrification in the fertilized tidal creek sediment was higher than in the control creek. However, no effect was found on anammox rates and anammox's contribution to the total N2-production remained insignificant under high nitrate conditions.
nitrogen balance in a large fluvial lake: denitrification, plant assimilation
and N fixation
Roxane Maranger, Laure Tall and Catherine Blanchet
Lake St Pierre is the last and largest fluvial lake in the St-Lawrence system and has been designated a UNESCO world biosphere site because of its extensive wetland habitat. The deep shipping channel (>13 m) that bisects this shallow lake (average depth 3 m) impacts hydrology creating 3 distinct water masses: 1) North mass draining the Ottawa River, 2) Central, the Great Lakes and 3) South, 3 tributaries heavily impacted by agriculture. We evaluated the relative importance of denitrification, N2O production and plant assimilation spatially in this system throughout summer 2005. Using a mass balance approach, the system retained 62 tonnes N-NO3+NO2 d-1, approximately 30% of the total amount of nitrate that was loaded. Approximately 8.4 tonnes N-DIN d-1 or 45% of what enter was retained in the South Mass where measured N2O flux varied both temporally, being highest early in the summer and spatially, gradually becoming relatively more important in the upper reaches later in the summer. By using a conservative N2O:N2 ratio, we estimate denitrification could account for the removal of 5 to 8.5 tonnes N d-1. However plant uptake was also very important and spatially heterogeneous in this section of the lake, being significantly higher in the upper than the lower reaches. This was not simply a function of greater plant biomass but due to 3 fold differences in the C/N ratio of the plants. Plants retained up to 5.2 tonnes N d-1. Efficient denitrification and plant assimilation resulted in severe N-limitation. To compensate, the South mass developed a dense and extensive (20 km2) benthic cyanobacterial biofilm bringing in an estimated 2-5 tonnes N d-1 from the atmosphere. Human alteration of riverine ecosystems that may be naturally suited to denitrification may become too efficient through excess nutrient load (both N and P) and changing hydrology resulting N deficits that cause the proliferation of N fixing cyanobacteria.
Working Abstract: I have begun a study of denitrification and trace gas emissions at Timberlake Farms, part of Great Dismal Swamp Mitigation Bank, a 1000 acre former pocosin/riverine swamp forest in Tyrrell County, in the Albemarle Sound basin in eastern North Carolina. It was historically timbered until 1973, when it was converted to a corn and soybean operation with drainage ditches and a pump to remove water from the fields. The pump is to be removed, and ditches strategically plugged, which will allow the property to be reflooded and will create a new stream/wetland complex across the site in November 2006. The experimental design, including locations of monitoring points across the expected flooding gradient, and pre-flooding data on soil parameters, denitrification potential, and trace gas emissions are presented.
uptake and denitrification rates determined from15N addition
experiments: Preliminary results from the LINX II study
Pat Mulholland, Bob Hall, Steve Hamilton, Bruce Peterson, Jen Tank, Linda Ashkenas, Lee Cooper, Cliff Dahm, Walter Dodds, Stuart Findlay, Stan Gregory, Nancy Grimm, Sherri Johnson, Bill McDowell, Judy Meyer, Geoff Poole, Maury Valett, Jack Webster [pdf]
Working Abstract: The Lotic Intersite Nitrogen eXperiment II (LINX II) study is a large, multi-investigator, intersite study of nitrate uptake rates and controls in streams draining watersheds of different land use types. We have developed and applied a field 15N tracer addition approach that directly measures in situ rates of total nitrate uptake and denitrification at the scale of entire stream reaches by determining longitudinal decline in tracer 15N-nitrate flux and appearance of tracer 15N in N2 and N2O downstream from tracer additions. We measured nitrate uptake and denitrification rates in headwater streams in 8 regions across the U.S. draining reference, agricultural, and urbanized catchments - a total of 72 streams. Preliminary results show that total nitrate uptake rates ranged from undetectable to 126 µg N/m2/s (maximum Vf of 0.03 cm/s) and denitrification rates from undetectable to 2.5 µg N/m2/s (maximum Vf of 0.01 cm/s). Denitrification consisted almost exclusively of N2 production and, on average, accounted for about 20% of total nitrate uptake. In general, denitrification rates were higher in urbanized and agricultural streams than in reference streams. The strongest predictors of total nitrate uptake rate were nitrate concentration and gross primary production rate (both positive). Denitrification rates were also positively related to streamwater nitrate concentrations. Results to date indicate that while denitrification is an important sink for nitrate in streams with high nitrate concentrations, it is usually considerably lower than nitrate uptake due to assimilatory processes.
in permeable continental shelf sediments on the South Atlantic Bight
Alexandra M. F. Rao and Richard A. Jahnke
Nitrogen cycling and total respiration rates have been measured seasonally in sediment columns packed with continental shelf sands from the USA South Atlantic Bight (SAB). Natural continental shelf seawater (unamended and spiked with Na2NO3) was continuously pumped through these high permeability (4.66 x 10-11 m2) sediments of low organic carbon (0.05%) and nitrogen (0.008%) content. Significant respiration (O2 consumption and ΣCO2 production) was observed in all columns, with highest respiration rates in summer. The majority (78-100%) of N regenerated in sediment columns with oxic bulk porewater was denitrified and measured as N2, while a small amount was released as NO3-. Despite the evidence of denitrification in oxic columns, a lag was observed before 29N2 and 30N2 production became detectable after 15NO3- addition, and outflow concentrations of these species were much lower than 28N2. These unexpected observations indicate that nitrification and denitrification are tightly coupled and somewhat isolated from bulk porewaters. Although respiration rates varied following the addition of 15NO3-, 29N2 and 30N2 production was always stimulated in columns with anoxic effluent, supporting the expected stimulation of heterotrophic denitrification in suboxic conditions. A rate of 3.8 - 19.3 x 1010 mol N yr-1, or 1.15 - 5.84 mmol N m-2 d-1 was calculated for benthic denitrification on the SAB, which can account for most or all new N inputs to this region. Metal and sulfate reduction were observed in columns with long residence times that developed anoxic outflow in summer and fall, when respiration rates were highest. These results suggest that denitrification in high permeability coastal sediments plays an important role in the global N cycle.
Although a tight coupling between nitrification and heterotrophic denitrification in suboxic microenvironments can explain observations in sediment columns with SAB sands, the nature of the connection remains unclear. Further attempts to characterize the dominant mechanism supporting N2 production in oxic SAB sands using a numerical model of solute transport and reaction in permeable sands with advective porewater flow and spherical microenvironments will be discussed.
Do urban wetlands
leak nitrogen? An analysis of controlling factors on denitrification
and nitrate leaching in an urban context
Emilie K. Stander and Joan G. Ehrenfeld
Wetlands are increasingly being used as management tools to combat the widespread problem of excess nitrogen in surface waters of the United States. This is particularly true in urban or urbanizing watersheds. However, due to extensive hydrologic alteration of wetlands located in urban landscapes, urban wetlands are often much drier than they have been historically. This may cause changes in nitrogen cycling processes, particularly denitrification. Urban wetlands may actually be acting as sources of nitrate to downstream water bodies rather than the sinks they are thought to be. To determine whether urban wetlands are leaking nitrogen, I studied hydrology and nitrogen cycling processes in fourteen palustrine, forested wetlands, all located within 50 miles of New York City. These wetlands represent a range of urban intensity in surrounding land use/land cover from high density, old urban landscapes to old field/newly suburbanizing landscapes. One autowell in each site recorded water table measurements four times daily. In situ rates of net nitrogen mineralization and nitrification were measured monthly during the same time period using the static core technique. Potential denitrification rates were measured monthly in laboratory incubations using the acetylene block technique. Throughfall collectors measured inorganic nitrogen inputs to nine of these sites, and lysimeters collected inorganic nitrogen outputs to groundwater at the same nine sites over a one year period. Hydrographs demonstrated that many sites have water table depths consistently below 30 cm (i.e., below the biologically active zone) for long periods of time. Many wetlands display uncharacteristically flashy hydrographs. Wetlands with dry or flashy hydrographs did have higher nitrification and lower denitrification rates than wetlands with more normal hydrology. Wetlands located in municipalities with higher population densities had higher concentrations of nitrate in throughfall; population density, however, was not a good predictor of nitrification or denitrification rates. There were no meaningful relationships found between nitrate inputs and outputs, or between nitrate outputs and nitrification or denitrification. There may be more nitrate retention in these systems than nitrification and denitrification rates would predict, possibly as a result of plant uptake.
denitrification flux over a whole drainage network
Marie Thouvenot, Gilles Billen, Josette Garnier
The objective of this study is to quantify the importance of denitrification in the removal of nitrogen within an entire river network. For this purpose, we linked a deterministic biogeochemical sediment model to a hydrological/biogeochemical model (Riverstrahler). The benthic model was able to represent the sediment early diagenesis and the nutrients exchange with the overlying water while the hydrological model dealt with transport and reaction of nitrogen within the water column. We first tested the benthic model independently of the Riverstrahler model using local measures of flux across the sediment-water interface. The model was able to represent ammonium, nitrate and oxygen fluxes similar to the measurements. We then coupled the benthic model to the drainage network model for an estimation of denitrification flux over the entire drainage network of the Seine river system.
A landscape based
approach to estimate riparian hydrological and nitrate removal functions
P. Vidon and A. R. Hill
This study evaluates a conceptual model developed for riparian zones in Ontario that links landscape hydrogeological characteristics to riparian groundwater hydrology and nitrate removal efficiency. Data from a range of riparian sites in the USA and Europe suggests that the riparian zone types identified in the model are consistent with patterns of riparian hydrology and nitrate flux and removal in many humid temperate landscapes. These data also support the view that a riparian width of < 20 m is often sufficient for effective nitrate removal unless riparian sediments are coarse-grained or nitrate transport occurs mainly in surface-fed groundwater seeps. This study assesses the possibility of using topographic, soil, surficial geology and vegetation maps to determine landscape attributes linked by the model to riparian zone hydrological functioning and nitrate removal efficiency. Although mappable data can help in determining broad classes of riparian zones, field visits are necessary to determine non-mappable riparian attributes such as seeps, organic horizons and permeable sediment depth in the riparian zone. This research suggests that the conceptual model could be used for landscape management purposes in most temperate landscapes with minor modifications and that the hydrological component of the model could be adapted for contaminants other than nitrate.
of aquatic denitrification in a 5th order river network
Wollheim, W.M., C.Vorosmarty, C. Hopkinson, B.J. Peterson
Aquatic systems can significantly influence export of nutrients to coastal zones. The strength and distribution of nutrient removal via denitrification within entire river networks are determined by interactions between surface water characteristics, river network geomorphology, and biological process rates. We used a river network N removal model to quantify nitrate removal by the total length of stream in each order class in the 5th order Ipswich River network in coastal Massachusetts. NO3-N is commonly 1-2 mg N / L in urbanized headwater streams in the basin. We applied the model using biological rate parameters determined from 15NO3 additions in the Ipswich basin (LINX2 experiments). We find that during low flow periods, denitrification accounted for 32% of nitrate inputs to aquatic systems. Water withdrawals are also a significant sink of nitrate, accounting for 20% of estimated inputs. Nitrate export was 48% of inputs, similar to the 53% observed. The role of both water withdrawals and biological removal processes declines during high flow periods. Higher order rivers were disproportionately important sites of nutrient removal due to the distribution of benthic surface area, determined by hydraulic and geomorphic factors. The results suggest that understanding whole river network N removal requires adequate representation of N cycle processes in both large and small rivers.